Whisper Transcriber Suite vs WhisperAI
What Whisper Transcriber Suite costs you is time on your own computer: on a Windows laptop with no NVIDIA graphics card, its Large models turn an hour of audio into a 90 to 150 minute job. We don't make the suite. WhisperAI is our own cloud service, there's nothing to install, and the same hour usually comes back in under a minute.
By Eman Khoubian, Founder & CEO
Is Whisper Transcriber Suite free?
Yes. It's free to download and use under the open-source BSD-3 license, with no account and no per-minute cost.
On a CPU, its README puts the Large models at 1.5 to 2.5 times a recording's length and Small at about half. An NVIDIA graphics card with CUDA is about 10 to 20 times faster. On M-series Macs, the project's site says Whisper runs on the graphics chip through Apple's MLX when the model has an MLX version. It gives no speed.
The suite also has a try page that runs in your browser with nothing installed. It takes files up to 10 minutes long (5 on a phone) in 16 languages, adds no speaker labels, and needs about as long as the audio itself. It works like Whisper Web.
If you saw a monthly price while searching, that was probably WhisperTranscribe, a different, paid service.
Whisper Transcriber Suite download: what to expect
The installers are on the project's GitHub. On Windows, you'll need:
- 64-bit Windows 10 or 11.
- 8 GB of RAM. Large v3 needs at least 6.5 GB.
Installer
About 251 MB
SmartScreen warning
Windows warns the build isn't code-signed
Speech model
0.5 to 3 GB, downloaded once
Transcribing
Large models: 1.5 to 2.5x the audio on a CPU
About 0.75 to 3.25 GB downloaded before your first transcript.
On a Mac, the builds aren't notarized, so macOS shows a similar warning. The Apple-silicon build is tested on GitHub's machines, not yet by hand on a real Apple-silicon Mac. Linux has no package, so you run it from source.
WhisperAI skips the first three of those steps. You open a browser tab and upload the file, up to 5 GB on every plan.
Is Whisper Transcriber Suite legit?
Yes. It's real open-source software, and you can read every line on GitHub. It's also new. The repository was created in May 2026, version 2.0 came out on October 10, 2026, and four more releases followed within 20 days. We found no independent reviews of it yet.
One setting is worth knowing. After each transcription, the desktop app sends usage statistics, on by default: the model, language, durations, word count, your system's details, your region and a random install ID, never the file's name. You can turn them off under Help, and File > Work offline stops all network use.
The name gets mixed up with others. Whishper is a different open-source project, now called Anysub, that runs on a Linux server with Docker. Whisper Transcription on the App Store is Good Snooze's app. Neither is ours. The suite runs OpenAI's open Whisper model, like other Whisper desktop apps. WhisperAI runs on AssemblyAI and isn't affiliated with OpenAI.
Whisper Transcriber Suite vs WhisperAI, side by side
Both export SRT, VTT, DOCX and PDF. Here's where they differ.
| Feature | WhisperAI | Whisper Transcriber Suite |
|---|---|---|
| Price | Free: 5 minutes a month, no card. Premium: $14.99 a month for 120 minutes. Business Pro: $24.99 a month or $199 a year, unlimited. | Free and open source (BSD-3) |
| Where the transcribing happens | On our servers, from a browser tab | On your computer |
| Speaker labels | Every plan, Auto or 2 to 10 speakers | On your computer, Auto or 2 to 10 speakers. Adds processing time. |
| Translation | Into 87 languages, on every plan | Into English only, and not with the default Large v3 Turbo model |
| Summaries | Business Pro | Off by default. Needs your own AI service and key, or a local model the installers don't include. |
| Recording live | A browser tab and your mic in Chrome or Edge, on every plan. Nothing joins the call. | Your mic, or system audio on Windows. No speaker labels. |
| From your phone | The iPhone app, or WhisperAI in any phone's browser | Send a file over your Wi-Fi to the computer running the app |